On the Behavior of FRP-to-concrete Adhesive Interface: Theoretical Models and Experimental Results

Enzo Martinelli, Antonio Bilotta, Ciro Faella and Emidio Nigro

Department of Civil Engineering, University of Salerno, Fisciano (SA), Department of Structural Engineering, University of Naples "Federico II", Naples

Italy

1. Introduction

Fiber-reinforced polymers (FRP) are more and more commonly employed for strengthening existing structures of both reinforced concrete (RC) and masonry. Since FRP sheets (cured in situ) or plates (preformed) are externally bonded on a concrete or masonry substrate, the issue of adhesion on those materials generally controls the effectiveness of strengthening in members stressed either in bending or shear (Motavalli & Czaderski, 2007).

The use of composite materials for structural strengthening of civil structures and infrastructures began with some pioneering application at the middle of the ’80s (Meier, 1987) of the past century.

Plenty of experimental work and theoretical investigations have been carried out in the following years with the aim of demonstrating the feasibility of strengthening civil structures by means of composite materials (Swamy et al., 1987; Meier, 1995). However, composite materials were already widely used in other fields of structural engineering, such as aerospace (Hart-Smith, 1973), aeronautics and, later, automotive. Thus, the initial research activities about the possible use of composites in civil structures were not mainly focused on the behavior of composites themselves. They were rather intended at addressing two main issues regarding, on the one hand, the different behavior of composites with respect to more traditional materials (basically, steel) commonly used as a reinforcement in civil structures (Arduini & Nanni, 1997; Naaman et al., 2001; Triantafillou et al., 2001) and, on the other hand, the aspects related to the adhesive connection of the FRP laminates to the concrete (or masonry) substrate (Taljsten, 1997; Neubauer & Rostasy, 1997).

The main findings of the research activities carried out in the ’90s contributed to guidelines (fib, 2001; CNR-DT200, 2004; ACI 440-2R-08) for designing FRP-based strengthening intervention of RC and masonry members.

Bonding between FRP laminates (sheets or plates) and concrete emerged as a cutting-edge issue from the first decade of research activities on composite materials for civil structures. In particular, several failure modes due to loss of adhesion between the externally bonded FRP element and the concrete substrate have been observed experimentally and recognized as specific features of this kind of members (Meier, 1995; Bonacci, 1996).

On the Behavior of FRP-to-concrete Adhesive Interface: Theoretical Models and Experimental ResultsAs a matter of principle, the following seven failure modes have been defined in the scientific literature (Teng et al., 2002):

a. flexural failure by FRP rupture;

b. flexural failure by crushing of compressive concrete;

c. shear failure;

d. concrete cover separation;

e. plate end interfacial debonding;

f. intermediate flexural crack induced interfacial debonding;

g. intermediate flexural-shear crack induced interfacial debonding.

Подпись: End debonding Подпись: Intermediate debonding Подпись: Debonding induced by a local irregularity Подпись: Shear crack-induced debonding

The last three failure modes are actually related to debonding failure of the FRP laminate from the concrete substrate. Local failure possibly induced by irregularities in the substrate surface can be also observed. Fig. 1 represents those failure modes pointing their typical position throughout the FRP-to-composite adhesive interface.

Fig. 1. Possible failure modes due to debonding in RC beams externally strengthened by FRP

Analytical studies have been carried out for determining the actual distribution of stresses throughout the adhesive interface. In particular, well-established elastic models, already used within the framework of structural mechanics, have been proposed as simplified practical methods for determining both shear and normal (peeling) stresses at the FRP-to – concrete interface (Roberts, 1989; Malek et al., 1998; Rabinovich & Frostig, 2000). However, those simplified methods were generally based on the assumption of an elastic behaviour of the above mentioned interface. Although this assumption can be considered under service loads, it cannot be generally accepted for the load values close to the onset of debonding. Micro-cracking phenomena develop as the levels of the interface stresses cannot be resisted by concrete, resulting in a highly nonlinear behaviour for the FRP-to-concrete interface which can be modelled by means of several possible relationships between interface stresses and displacements (Lu et al, 2005). Consequently, more advanced numerical models have been also proposed for simulating the actual distribution of stresses looking after the various fracture phenomena developing in concrete beneath the adhesive interface (Rabinovich & Frostig, 2000; Faella et al., 2008).

More recently, several models have been proposed in the scientific literature for predicting the strength of beams against both plate end-debonding and intermediate-debonding (Smith & Teng, 2001; Teng et al., 2003). They generally derive by mechanical observations carried out on the behaviour of FRP laminates glued on concrete blocks (Taljsten, 1996) and tested in pullout with the aim of measuring some relevant quantities, such as the ultimate strength
at debonding and the axial strain distribution throughout the FRP bonded length (Chajes et al., 1996).

Understanding the behaviour of FRP-to-concrete joints tested under pull-out actions is of paramount importance for describing the key mechanical properties of the adhesive interface between FRP and concrete, which plays a key role in the possible debonding failure of externally strengthened beams. Thus, advanced testing and monitoring techniques have been also used for a deeper investigation of the behaviour of those joints (Czaderski et al., 2010). The experimental results obtained by the mentioned pull-out tests can be considered for identifying the non-linear relationships connecting interface stresses and displacements (Faella et al., 2009). In particular, a bilinear elastic-softening relationship between the interface shear stress and the corresponding displacements is often used for the FRP-to-concrete interface. The elastic branch of that stress-strain relationship results from the elastic behaviour of both the adhesive layer and the concrete substrate, mainly stressed in shear. The slope of such an elastic branch (namely, the "slip modulus" according to Lee et al, 1999) is generally much smaller than the value corresponding to the ration between the shear modulus of the resin and its thickness, as it would be determined by assuming a fully stiff behaviour of the concrete block (Faella et al., 2002).

A closed-form analytical solution has been derived in Faella et al. (2002) for the distribution of both shear stresses and relative slips in FRP-to-concrete joints with a bilinear adhesive interface. Further advances have been proposed by Lu et al. (2005), while the influence of different assumptions on the shape of the stress-slip relationship (i. e. bilinear, linearly softening, rigid-plastic, and so on) is discussed in both Chen & Teng (2001) and Wu et al. (2002). However, the fracture energy GF (Taljsten, 1996) is the key parameter characterizing any various shear-stress-interface-slip relationship. The ultimate pull-out strength of the FRP-to-concrete joints is controlled by that parameter as well as the axial stiffness of the laminate and the bonding length. A limit value of the bonding length beyond which no further increases of the ultimate pull-out force can be observed; it is generally referred as "transfer length" (Bizindavyi & Neale, 1999). Some concepts of Fracture Mechanics are more and more employed in modelling the overall behaviour of both FRP laminates connected to concrete blocks (Yuan et al., 2007) and RC beams externally strengthened by FRP (Rabinovitch & Frostig, 2001; Achintha & Burgoygne, 2008).

The present chapter deals with the mechanical behaviour of the adhesive interface between FRP and concrete, proposing a wide overview of some theoretical models and experimental results which can be useful for understanding such a behaviour.

In the second section, an analytical model will be firstly presented for determining both shear and normal stresses throughout the adhesive interface in the linear range. The nonlinear behaviour of the FRP-to-concrete interface will be also addressed by discussing the ultimate bearing capacity of FRP laminates bonded on concrete members.

The third section will present the features of the most well-established testing techniques for investigating the mechanical behaviour of FRP-to-concrete interface, from the early loading stages up to failure. A series of experimental results of pull-out tests on FRP laminates will be also presented for pointing out the most important behavioural characteristics of the FRP-to-concrete interface in the case of both sheets and plates.

Finally, the fourth section will present some of the most recent design formulae which can be derived by the above mentioned theoretical models and calibrated on the available experimental results, partly reported in the present chapter.